Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 Brang¨ane:a new family of Barbarian

A. Cellino1, S. Bagnulo2, P. Tanga3, M. Devog`ele4, Ph. Bendjoya3, E. Reilly2, J.-P. Rivet3 and F. Spoto3,5 1 INAF - Osservatorio Astrofisico di Torino, Pino Torinese 10025, Italy. E-mail: [email protected] 2 Armagh Observatory, College Hill, Armagh BT61 9DG, UK. 3 Universit´ede la Cˆoted’Azur, Observatoire de la Cˆoted’Azur, CNRS, Laboratoire Lagrange, France 4 Lowell Observatory, 1400 West Mars Hill Road, Flagstaff, AZ 86001, USA 5 IMCCE, Observatoire de Paris, 77 av. Denfert-Rochereau, 75014 Paris, France

Accepted After. Received Before; in original form: Earlier

ABSTRACT The so-called Barbarian asteroids (from the name of the prototype of this class, as- teroid (234) Barbara) are extremely interesting objects because they might be the survivors of a very early generation of planetesimals. We have identified in the past the first case of an family (the Watsonia family) whose members are Barbar- ians, all issued from the collisional disruption of a sizeable parent body. Here we report on the identification of another family, called after the name of its biggest member, asteroid (606) Brang¨ane,consisting of objects displaying the polarimetric properties characterizing the Barbarian behaviour. This is the first recognized case of a swarm of small Barbarians issued from a quite recent cratering event. Key words: asteroids: polarization – asteroids: composition.

1 INTRODUCTION metric data, has been recently published by Devog`eleet al. (2018). These authors found conclusive evidence that all The discovery of an anomalous polarimetric behaviour of Barbarians, which had been previously classified as mem- (234) Barbara (Cellino et al. 2006) was the first recognition bers of different taxonomic classes based on their reflectance of the existence of a new class of asteroids (now commonly spectra at visible wavelengths, once the properties of both referred to as the Barbarians) which are very interesting the visible and near IR regions of the reflectance spectrum in many respects. Some of these objects are known to ex- are taken into account, appear to belong to a unique taxo- hibit anomalous reflectance spectra, characterized by fea- nomic class: the L class defined by DeMeo et al. (2009). tures that were interpreted by Burbine et al. (1992); - Devog`eleet al. (2018) found that the spectra of L-class shine et al. (2008) as diagnostic of a very high content of objects can be successfully modeled using mixtures of the refractory compounds on their surfaces, including primar- fluffy kind of spinel present in CAIs, MgO-rich olivine and ily the aluminous spinel mineral (MgAl2O4). This is a ma- the mineral compounds found in the matrix of CV3 mete- jor component of the so-called Calcium Aluminum-rich In- orites. They also found that the presence of CAIs is respon- clusions (CAI) found in some meteorites. With their ages sible for the anomalous polarimetric inversion angle char- around 4.567 Gyrs, derived from isotopic abundances, CAIs acterizing the Barbarians (for an explanation of this and are the oldest samples of solid matter found in the Solar sys- other polarimetric parameters, see, for instance, Cellino et tem. As pointed out by Sunshine et al. (2008), CAIs are gen- al. 2015). In particular, they pointed out the presence of a erally classified, on the basis of petrography and geochem- variation of the inversion angle of asteroid (234) Barbara istry, into three major groups. Among them, the so-called as a function of wavelength, and interpreted it as a conse- ”fluffy type A” (FTA), characterized by a porous, strongly quence of the known wavelength-dependent variation of the non-compact structure, is found in all chondritic meteorites, refractive index of the spinel mineral. and produces a strong absorption feature in the reflectance Until a couple of years ago, it was not clear whether the spectrum, interpreted as a consequence of a high concentra- properties of Barbarians were determined by anomalous sur- tion of FeO-rich spinel. It is also believed that, as opposite to face properties, or by structural composition. In this respect, the other types of CAIs, FTAs may never have been totally a major step forward has been the discovery by means of po- melted by any transient heating event after their formation. larimetric measurements that the Watsonia The most extensive analysis of objects belonging to the (Novakovi´cet al. 2011) consists of Barbarian objects (Cellino Barbarian class, based on both spectroscopic and polari- et al. 2014). Table 1. Total number of currently recognized members of the Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 Brang¨anefamily, and the lower and upper limits of the intervals covered by its members in the space of the proper semi-major axis, eccentricity and sine of inclination.

N. of proper semi-major proper sine of proper members axis (AU) eccentricity inclination

325 2.571 – 2.597 0.178 – 0.183 0.165 – 0.168

Figure 2. The same as Fig. 1, but for the family of Watsonia, the only one previously identified family of Barbarians (Cellino et al. 2014). As shown by the morphology of its cumulative size distribution, the Watsonia family may suggest a more energetic disruption event, rather than a typical cratering event as in the case of the Brang¨anefamily.

been taken from the AstDys site1, where a regularly updated list of members of asteroid families is publicly available. The cumulative size distribution of the family is shown in Fig. 1. The sizes of the family members have been taken directly from the NEOWISE catalog (Mainzer at al. 2016), whenever available. The same data have been also Figure 1. Cumulative size distribution of the Brang¨anefamily. used to derive the average albedo of the family, 0.116 ± The typical trend characterized by a large difference in size be- 0.005, very similar to the of Brang¨aneitself tween the largest member and the rest of the family members (0.113±0.021). The sizes of family members not observed by clearly shows that the family is the outcome of a cratering event. WISE were computed from the known relation linking the At sizes smaller than about 2 km the size distribution becomes size to the absolute magnitude H and the geometric albedo shallower, an indication of incompleteness of the recognized small- est members pV :

log (D) = 3, 1236 − 0.2 H − 0.5 log (pV ) (1) where D is the equivalent diameter in km. In the compu- tations, the value of H was taken from the file listing the Asteroid families are groups of objects issued by the family members available in the AstDys site. For the value collisional disruption of single parent bodies. The fact that of the geometric albedo p we adopted the NEOWISE based the members of the Watsonia family are Barbarians strongly V average value of 0.116 mentioned above. suggests that the Barbarian properties are not purely surface The Brang¨anefamily is the likely outcome of a crater- effects, because the members of a same family derive from ing event. This is suggested by the large difference in size material located at different depths inside the parent body. between its largest member, (606) Brang¨ane,having an ab- This was also the first discovery of Barbarian asteroids of solute magnitude H = 10.22 and an estimated size of about relatively small sizes. 36 km (Mainzer at al. 2016), and the rest of the family mem- The main belt asteroid (606) Brang¨ane,about 40 km in bers. The latter have absolute magnitudes generally fainter size (Mainzer at al. 2016), belongs to the L-class (DeMeo et than 14, and sizes mostly below 5 km. In the cases of families al. 2009). It is the biggest member of another family (Milani issued from cratering events, the overall trend of the cumu- et al. 2014). Table 1 summarizes the main parameters of the lative size distribution tends to exhibit at the largest sizes Brang¨anefamily which, according to Spoto et al. (2015) is the result of a very recent cratering event (the age between 30 and 50 Myrs being derived by the analysis of the “V- shaped” diagram, see below) In this Table, the data have 1 http://hamilton.dm.unipi.it/astdys/index.php?pc=5 a characteristic concave trend (Tanga et al. 1999; Durda et Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 al. 2007). For a comparison, Fig. 2 shows the cumulative size dis- tribution of the Watsonia family. The morphology of the size distribution, computed in exactly the same way as in the case of Brang¨ane,as explained above, is clearly different with respect to the case of Brang¨ane,as it is not charac- terized by a big difference between the size of the largest member and those of the other biggest family members. It can also be noted that in the case of Watsonia there is a stronger depletion of objects at the low-end of the size dis- tribution. This can be partly due to observational bias, due to the fact that the perihelion distances of the Brangane family are lower than those of Watsonia members, due to a non-negligible difference in . On the other hand, the two families are located at fairly similar heliocen- tric distances (the orbital semi-major axis of (729) Watso- nia is about 2.76 AU) and, consisting of objects displaying similar spectroscopic and polarimetric properties, also the albedo of the members of the two families should be simi- lar. Our preferred interpretation is that Watsonia represents the outcome of a collision producing a more extensive dis- ruption of the parent body, and the relative lack of small family members with respect to the case of the Brang¨ane Figure 3. Polarization versus phase-angle data for the targets of the present investigations, as well as a couple of measurements family could be diagnostic of an older age, because the pro- of (606) Brang¨ane,the largest member of the family, previously gressive depletion of the smallest members of a family due to observed at the Calern observatory. The family data are compared the Yarkovsky effect increases with time (Rubincam 1995; with the polarization curve of the prototype of Barbarians, (234) Farinella et al. 1998). This interpretation is also supported Barbara, whose available measurements are shown, for sake of by the apparently shallower slope of the cumulative size dis- clarity, without their associated error bars (orange symbols). tribution of the Watsonia family if we compare it with that of the Brang¨anefamily. A progressive decrease of slope as a function of time is expected to be a typical result of a PQ = Pr, and for symmetry reasons we expect PU to be Yarkovsky-driven erosion. consistent with zero. In fact, PU measurements may be used The reason why the Brang¨anefamily is so interesting for quality check. If the light is polarised in the direction is that we recently discovered that (606) Brang¨ane itself is perpendicular to the scattering plane, then Pr = PQ > 0, a Barbarian (Devog`eleet al. 2018). It was therefore very while if the light is polarised in the direction parallel to the tempting to carry out a campaign of polarimetric observa- scattering plane, then Pr = PQ < 0. The latter situation tions to check whether we could find a new example of a is generally described with the potentially misleading term family consisting of Barbarian objects. The results of this of negative polarisation. Practically one finds that at small investigation are presented in this paper. phase angles ( 20◦) all asteroids exhibit negative polar- ◦ isation, while at phase angles (& 20 ) nearly all asteroids exhibit positive polarization. Barbarian asteroids represent an exception because their polarisation changes sign around 2 NEW OBSERVATIONS ◦ phase-angle & 27 ÷ 30 . To identify Barbarians, we decided Our target list included ten objects that are members of the to observe our targets at phase-angles in the range 17 - 30◦, Brang¨anefamily according to the list of family members and establish whether the polarization would change sign available at the AstDys web site. These objects are listed in around 18-22◦ or around 27–30◦. Table 2. We observed our targets using the VLT FORS2 in- The light that we receive at visible wavelengths from strument (Appenzeller et al. 1998) in imaging polarimetric the asteroids is scattered sunlight, in a state of partial lin- mode, and obtained 15 broadband linear polarization mea- ear polarization, either parallel or perpendicular to the Sun surements in the R special filter from April to September - target - observer plane (the scattering plane). Because the 2017. Polarimetric measurements were performed with the direction of the polarisation is identified by the view geom- retarder wave-plate at all positions between 0◦ and 157.5◦, etry, in the solar system science literature the state of linear at 22.5◦ steps. For each observation, the exposure time ac- polarization is traditionally described by just one parame- cumulated over all exposures varied from 720 s (for 87168) ter, Pr, defined as the difference between flux perpendicular to 2857 s (for 76378). Polarimetric data were treated as ex- to the scattering plane and the flux parallel to the scat- plained in Bagnulo et al. (2016); observing log and results tering plane, divided by their sum. In more general cases, are given in Table 2. linear polarisation is described by the two reduced Stokes Each polarimetric series was preceded by the acquisition parameters PQ = Q/I and PU = U/I. By adopting the of an image to fine-tune the telescope pointing and ensure direction perpendicular to the scattering plane as a refer- target is properly centred and not vignetted by the Wollas- ence direction for the Stokes parameters Q and U, we have ton masks. These acquisition images were used for photo- Table 2. and photometry of ten asteroids of in the special R filter. PQ and PU are the reduced Stokes parameters measured Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 in a reference system such that PQ is the flux perpendicular to the plane Sun-Object-Earth (the scattering plane) minus the flux parallel to that plane, divided by the sum of the two fluxes. In literature PQ is usually designated with the symbol Pr. In the last columns, mR(α) and m(α) are the magnitude reduced to unit distance from the observer and from the Sun, and the apparent magnitude, respectively.

Date Time (UT) Exp Object Phase angle PQ PU (yyyy mm dd) (hh:mm) (sec) (DEG) (%) (%) mR(α) m(α)

2017-07-04 09:53 904 36358 30.00 0.82 ± 0.11 0.03 ± 0.11 16.07 ± 0.20 18.88 ± 0.20 2017-08-01 06:16 904 28.66 0.67 ± 0.11 0.06 ± 0.11 16.22 ± 0.21 18.71 ± 0.21 2017-08-17 09:29 496 26.02 0.02 ± 0.11 0.03 ± 0.11 16.24 ± 0.20 18.53 ± 0.20 2017-04-26 09:28 824 51187 22.04 −0.84 ± 0.17 0.13 ± 0.17 20.25 ± 0.13 22.98 ± 0.13 2017-07-28 02:01 860 17.34 −1.56 ± 0.10 0.06 ± 0.10 18.48 ± 0.52 21.31 ± 0.52 2017-06-19 01:14 992 53707 24.33 −0.35 ± 0.12 0.18 ± 0.12 16.08 ± 0.16 19.23 ± 0.16 2017-07-24 23:45 1568 26.29 0.12 ± 0.13 0.02 ± 0.13 16.01 ± 0.19 19.48 ± 0.19 2017-09-20 04:40 1568 56748 17.76 −1.36 ± 0.07 0.17 ± 0.07 16.38 ± 0.28 19.00 ± 0.28 2017-06-25 01:01 1496 75890 17.24 −1.48 ± 0.08 −0.01 ± 0.08 16.40 ± 0.12 19.14 ± 0.12 2017-07-17 02:36 2778 22.25 −0.79 ± 0.14 −0.06 ± 0.14 16.41 ± 0.11 19.49 ± 0.11 2017-06-21 04:34 1368 76378 17.54 −1.60 ± 0.08 −0.08 ± 0.08 16.31 ± 0.11 18.69 ± 0.11 2017-07-17 03:33 2856 24.18 −0.75 ± 0.11 0.20 ± 0.11 16.73 ± 0.28 19.50 ± 0.11 2017-06-23 00:30 1136 82955 24.25 −0.50 ± 0.08 0.05 ± 0.08 16.30 ± 0.11 18.86 ± 0.11 2017-08-16 01:02 2376 27.95 0.48 ± 0.14 0.48 ± 0.14 16.43 ± 0.13 19.62 ± 0.13 2017-08-28 09:26 720 87168 27.26 −0.08 ± 0.12 0.07 ± 0.12 16.78 ± 0.16 18.97 ± 0.16 2017-06-22 23:46 2072 103450 26.26 0.23 ± 0.11 0.13 ± 0.11 16.53 ± 0.19 19.74 ± 0.19 2017-09-12 03:23 2376 116799 20.55 −0.86 ± 0.15 0.06 ± 0.15 16.61 ± 0.06 19.48 ± 0.06

metric measurements. PSF fitting photometry of all objects Table 3. New polarimetric measurements of (234) Barbara in the in the images was performed using the procedures in the V filter. Uncertainties are somewhat arbitrarily estimated based DAOPHOT package of iraf using the following strategy. Ob- on our knowledge of the instrument polarization. Contribution jects where a maximum of ten iterations were required in or- from photon-noise is . 0.01 %. der to fit the PSF and no error was recorded for the resulting photometry were kept for the following steps. The objects were matched by sky coordinate to those recorded in the Observation date Phase angle PQ PanSTARRS Mean Object Catalog. For each image, a lin- (deg) (%) ear regression model was fit for the measured apparent mag- 2015-05-09 13.5 −1.54 ± 0.05 nitude mr, to the Mean PSF magnitude, mr(PanSTARRS), 2015-05-27 18.5 −1.33 ± 0.05 recorded by PanSTARRS in the r filter. An average of 100 2016-07-07 33.2 0.32 ± 0.05 objects per image were used to generate the model. The model was then used to predict the magnitude of the target asteroid. Reduced magnitudes were computed using helio- and geo-centric distances taken from HORIZONS (Giorgini et al. 1996), and are given in Table 2, together with the cor- responding apparent magnitudes. It is important to com- It is also useful to point out that FORS acquired polari- ment that, at least in the present case, we do not expect metric standard stars within its calibration plan, and the that individual photometric measurements can provide the fact that PU is always consistent with zero, means that the same physical insight as polarimetric measurements, as the polarimetric optics were consistently well aligned. limited phase-angle range covered by our observations does Our polarimetric measurements are shown in Figure 3, not even allow us to estimate the slope of the magnitude- in which we show also two measurements of (606) Brang¨ane phase angle curve. However, both photometry and polarime- obtained in the V filter by Devog`eleet al. (2018). A compar- try data could be useful in a future work for modelling pur- ison is also made with the phase-polarization curve of (234) poses, if more data are collected. Barbara, the prototype of this class of objects using liter- It is important to note also that the errors on the polari- ature data (Cellino et al. 2016; Devog`eleet al. 2018). and metric measurements were calculated based only on photon adding the three new measurements of Table 3 obtained in noise using normal error propagation of the flux errors (see, the framework of a recent spectro-polarimetric survey of as- e.g., Bagnulo et al. 2009). Systematics with FORS2 are of teroids (observations and data reduction were performed in the order of a few units in 10−4 (e.g., Bagnulo et al. 2016). the same way as Bagnulo et al. 2015). 3 DISCUSSION AND CONCLUSIONS a formation dating back to epochs when the solid matter Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 accreted into growing planetesimals was characterized by a All asteroids of our target list exhibit the Barbarian polari- much higher abundance of refractory components, the first metric behaviour, characterized by a wide negative polar- compounds to solidify during the very early cooling of a ization branch, and an inversion angle which in this case is ◦ hot proto-planetary disk. Barbarians could therefore be very ∼ 25 . The two independent CAPS (Calern Asteroid Polari- old, having possibly survived the very chaotic early phases metric Survey) measurements of (606) Brang¨aneperfectly fit of evolution of our planetary system. the trend exhibited by the other family members. According to the so-called “Grand Tack” model (Walsh A more detailed analysis of Fig. 3 shows that the et al. 2011, 2012), the very early history of the Solar system Brang¨anefamily members very closely mimic the behaviour was characterized by episodes of back and forth migration ◦ of (234) Barbara in an interval of phase angles ∼ 17 ÷ ∼ of the giant planets, producing a huge depletion (by more ◦ 25 . At the highest values of phase angle, however, we note than 99 %) of the planetesimals originally accreted in the that the behaviour of (234) Barbara is not clearly defined be- region of the current main . The subsequent ◦ cause the four measurements at phase-angle & 30 are not in evolution, up to the end of the so-called Late Heavy Bom- a good agreement among each other. In particular, the mea- bardment phase, about 3.9 Gyrs ago, is described by the ◦ surement obtained at a phase of 30.4 obtained by Cellino et so-called “Nice” Model (Gomes et al. 2005; Morbidelli et al. al. (2016) fits very well the data obtained for the members 2005). It is believed that this was also an era of instability for of the Brang¨anefamily, but is high when compared with the small bodies, subject to perturbations that led many ob- the other measurements obtained at similar or near phase jects to move from the outer regions to smaller heliocentric angles, which suggest a much gentler increase of polariza- distances. At the end of this era, the major planets achieved tion in the positive branch. More observations are needed to their current configuration, and the populations of small better characterize the behaviour of (234) Barbara around bodies orbiting at different heliocentric distances started to ◦ phase angle 30 . Based on the observations presently avail- be more stable, their further evolution being mainly deter- ◦ able, the inversion angle could be from 25 (very similar to mined by mutual collisions and, in the current asteroid belt, ◦ the inversion angle for the Brang¨anefamily) to about 30 . by a steady removal of the smallest objects due to a pro- In any case it is evident that our measurements of a gressive drift of their orbital semi-major axes determined sample of members of the Brang¨anefamily definitively show by thermal emission mechanisms (Yarkovsky effect, see, for that this is a family of Barbarians, and the observed be- instance, Rubincam (1995); Farinella et al. (1998)). So, the haviour of its members suggest an inversion angle of polar- lucky survivors of an earlier age are expected to be rare, and ization between 25◦ and 27◦. Interestingly enough, according larger than a few tens of km. to Devog`eleet al. (2018), such a value would suggest a rel- This could explain why Barbarians are so rare and rel- atively low abundance of CAI abundances, possibly smaller atively big: apart from the members of the Watsonia and than 5 %. Moreover, although our data do not allow us to Brang¨anefamilies, there are currently only 17 identified Bar- derive a robust estimate of the extreme value of negative barians in the whole main belt, and all of them are above polarization Pmin, the visible trend seems to suggest a fairly 40-50 km in size. Smaller objects could have avoided detec- deep value for this parameter. According to Devog`eleet al. tion so far simply because they are fainter, but if they were (2018) this would also correspond to small abundances of common it is unlikely that a large number of them would nano-phase iron particles on the surface (see Fig. 13 of De- not have been identified in spectroscopic surveys like the vog`eleet al. 2018). Further spectroscopic data in the visible one carried out by DeMeo et al. (2009), taking into account and near-IR spectral regions will help to confirm or rule out that the L taxonomic class is characterized by some evident these tentative inferences. features in the near-IR region of the reflectance spectrum. The existence of a variety of values for the inversion an- Among the few known Barbarians, moreover, there are gle of polarization of different Barbarians can be important, reasons to believe that four objects, namely (387) Aquitania, and may correspond to significant differences in terms of sur- (729) Watsonia and (980) Anacostia, as suggested by Cellino face properties and composition, according to preliminarily et al. (2014), and also (599) Luisa, found more recently to be inferences by Devog`eleet al. (2018). This means that a pos- another nearby Barbarian, could be the surviving members sible difference of about five degrees between the inversion of a previous, first-generation family produced by the colli- angle of the Brang¨anefamily and (234) Barbara suggests sional disruption of a much bigger Barbarian parent body, that the small population of objects that we call Barbar- occurred presumably in very ancient times. The idea is that ians might be more heterogeneous than currently believed. the parent body of the Watsonia family was likely a fragment As shown by Devog`eleet al. (2018), the reflectance spectra from an earlier, first-generation collision involving a much of different Barbarians may also exhibit differences that in bigger parent body. According to Cellino et al. (2014), this some cases might represent more than subtle nuances among event produced, in addition to the parent body of the Watso- objects sharing a given set of properties. In other words, the nia family, also a few other big Barbarian asteroids (includ- data currently available suggest that the Barbarians might ing (387) Aquitania and (980) Anacostia, see also Masiero include objects having non-negligible differences in terms of & Cellino (2009)). After a very long time since their origin, surface properties and composition. Only by obtaining new these objects have orbits still similar to each other and to data we can confirm or rule out this possibility. those of the Watsonia family, but not sufficiently close to Devog`ele et al. (2018) claimed that the polarimetric each other and to Watsonia to form an identifiable, unique properties of Barbarians are a consequence of the high con- family. In other words, according to Cellino et al. (2014) we tent of spinel-bearing minerals in their surfaces, confirming would have here the first recognized example of an eroded previous results by Sunshine et al. (2008). This may suggest and dispersed family that was formed during or just later than the early, chaotic phases of the solar system history, that Barbarians are characterized by a composition rich in Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 characterized by strong perturbations due to processes of fluffy-type CAIs. The production of this mineral assemblage migration of the giant planets. The few survivors of this seems to require a mild, but not negligible degree of ther- very old family are still recognizable today as siblings shar- mal alteration, whereas a strong thermal alteration would ing a same origin, only as a consequence of their peculiar be expected to destroy the fluffy-type, iron-enriched vari- physical properties. ety of spinel. The interior of planetesimals big enough to The age of this first-generation family that supposedly include a significantly large amount of Al26 seems not to be produced (729) Watsonia and the big Barbarian asteroids a favorable environment. In this respect, the existence of a nearby, must be older than the upper limit found by Spoto big Barbarian parent body from which a first generation of et al. (2015) for the ages of families still recognizable today fragments including (729) Watsonia and its nearby Barbar- using current techniques of family identification and dating. ian siblings (Aquitania, Anacostia, etc.) originated can be a The techniques of family identification are purely based on problem, since the required size of such parent body cannot the vicinity in the proper elements space, while the ages are be less than 200 km. Of course, more detailed models of the derived from the measurement of a size-dependent spread- internal structure and evolution of such a body is needed ing in proper semi-major axis of family members due to the to confirm or rule out the compatibility with the presence Yarkovsky effect, derived from analysis of the so-called “V- and survival up to our days of large amounts of fluffy-type, shaped” diagrams of families2. In other words, the ages are spinel rich CAI materials. estimated by means of an analysis of the time needed to In general terms, a composition enriched in refractory produce the observed, size-dependent spreading in orbital minerals can be more compatible with an origin in a hot semi-major axis of the family members, as a consequence of environment at moderate heliocentric distances, because in the Yarkovsky effect (Spoto et al. 2015; Milani et al. 2017). a colder environment a larger variety of minerals can exist in In the case of the Watsonia family, an estimate of the age the solid state, making it more difficult to explain a strong is uncertain, due to the asymmetric structure of the family. enrichment in refractory compounds. The age turns out to range between about 0.8 and 1.2 Gyrs, The determination of the most likely astrophysical en- depending on the unknown density of its members. Such vironment to explain the origin of Barbarians is an open an age, in spite of the large uncertainty, supports the as- question, due to our insufficient understanding of the com- sumption that the supposedly big primordial family that position of these objects. Devog`eleet al. (2018) obtained produced the Watsonia parent body could have an age of reasonable fits of the observed spectroscopic properties of the order of, if not larger than, 2 · 109 years, older than any Barbarians by considering a mineral assemblage composed other known family. by a mixture of fluffy CAI, MgO-rich olivine and minerals If the few existing Barbarians (apart from the small present in the matrix of a couple of CV3 meteorites. The members of more recent families) are long-sought survivors CV3 are low-albedo, primitive meteorites, and for some of from a very early , we might be close to have found them the presence of aqueous alteration cannot be ruled some pieces of the Holy Grail in asteroid science. In this re- out, something that is in tension with an origin in a hot spect, many primitive, low-albedo asteroids that are abun- environment. Mg-rich silicates, however, are predicted by dant in the current asteroid main belt are thought to have thermodynamical models to condense in a hot gas at tem- been originally accreted at much higher heliocentric dis- perature above 1200 K, whereas the reaction with Fe occurs tances (Gomes et al. 2005; Morbidelli et al. 2005). Recently, at lower temperatures. According to Hanner and Bradley Delb`oet al. (2017) found indications of the existence in the (2004), Mg-rich olivine (forsterite) and pyroxene (enstatite) inner asteroid belt of low-albedo objects that might be the may have formed by direct condensation in the inner pri- survivors of at least one asteroid family possibly originated mordial nebula, although both forsterite and enstatite have as early as 4 billion years ago. It is not clear whether the been found by the Giotto and Vega missions to be present in supposed parent body of this family can be representative 1P/Halley (Schultze et al. 1997), and, more recently, of a population of dark planetesimals accreted at relatively in Hale-Bopp (Crovisier et al. 1997) and 9P/Tempel small heliocentric distances, or might have been a member of (Lisse et al. 2007), based on thermal IR spectra. They have a migrating population of planetesimals originally accreted been also directly found in samples of comet 81P/Wild2 col- at larger heliocentric distances. lected by the Stardust mission (Clemett et al. 2007). However, the Barbarians look different with respect to All this suggests that we cannot rule out the possibility low-albedo objects, whatever their origin might be. In par- that Barbarians may be the few survivors of a pristine popu- ticular, Barbarians seem to have significantly higher lation of moderate-sized, early planetesimals accreted at the (mostly between about 0.15 and 0.20, according to Cellino beginning of planetary formation at moderate heliocentric et al. 2016), and are found in the current asteroid main belt distances. The situation, however, is still strongly ambigu- in a wide interval of heliocentric distances. ous and we have not yet a sufficient body of evidence to Devog`ele et al. (2018) found that the 2.5 µm absorption allow us to draw firm conclusions about the origin of these band which characterizes the reflectance spectra of Barbar- bodies, and we cannot yet discriminate among very different, ians seems to be diagnostic of the presence of iron-enriched possible scenarios. spinel mineral present in CAIs. In particular, they proposed The identification of a new family of Barbarians is in any case an important result. The total number of iden- tified Barbarians increases by many tens of objects, if we 2 Plots of the V-shaped diagrams of asteroid families can be assume that, apart from some fraction of random interlop- displayed using the Plotting the asteroids facility available at ers (Migliorini et al. 1995), most members of the Brang¨ane http://hamilton.dm.unipi.it/astdys2/Plot/ family, including those that have not yet been observed, are Barbarians. On the other hand, all the members of any ACKNOWLEDGMENTS Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 given family (apart from possible special cases of mutually- This survey is based on observations collected at the overlapping multi-generation events) are fragments pro- European Organisation for Astronomical Research in the duced by the disruption of one single parent body. This Southern Hemisphere under ESO programme 099.C-0206 means that all the members of a given family normally cor- (PI=Cellino). New polarimetric measurements of asteroid respond to one single asteroid of a previous generation. The (234) Barbara were obtained under ESO programmes 095.C- Watsonia and Brang¨anefamily members are fragments of 0925 and 097.C-0853 (PI=Bagnulo). New polarimetric mea- two, larger Barbarians that were disrupted in relatively re- surements of asteroid (606) Brang¨anewere obtained us- cent times. In particular, there is evidence that the Brang¨ane ing the Torino Polarimeter attached to the 1-meter ’Omi- family is very young Spoto et al. (2015). cron’ telescope of the C2PU facility (Observatoire de la Cte d’Azur, Calern site), in the framework of the Calern Aster- It is worth noting that the two impacts that pro- oid Polarimetric Survey (CAPS) carried out in collaboration duced the Brangane and Watsonia families did not cause a between the INAF - Torino Astrophysical Observatory and widespread alteration of the polarimetric properties of their the Observatoire de la Cˆoted’Azur (Nice, France). members, despite the fact that one was a catastrophic event and the other a cratering (which would create fresh-looking ejecta but leave a largest remnant having on the average REFERENCES a much older surface). This may imply that the long in- version angle property that characterizes Barbarians is not Appenzeller I. et al., 1998, The Messenger, 94, 1 affected by compaction, weathering, regolith size, or other Bagnulo S. et al., 2016, A&A, 585, A122 time-dependent effects, in agreement with results of other Bagnulo, S., Cellino, A., Sterzik, M.J., 2015, MNRAS, 446, studies. L11 Bagnulo S., Belskaya I. N., Boehnhardt H., Kolokolova L., Muinonen K., Sterzik M., Tozzi G.-P., 2011, JQSRT, 112, The recent production of small Barbarian fragments in 2059 the asteroid main belt is interesting, because it means that Bagnulo S. et al., 2009, PASP, 121, 993 at least some of these objects can have chaotically evolved, being subject to the Yarkovsky effect, up to possibly achiev- Bottke W. F. et al., 2015, in Asteroids IV (P.Michel, F. E. ing near-Earth asteroid-like orbits. As a consequence, it is DeMeo, W. F. Bottke, Eds.), 895, Univ. of Arizona Press, possible that at least a few Barbarians might exist today Tucson among the NEA population. This is exciting, because the Burbine T. H., Gaffey M. J., Bell J. F., 1992, Metic, 27, identification of some Barbarian NEAs would possibly per- 424 mit to conceive space missions of exploration and sample- Cellino A., Bagnulo S., Belskaya I.N., Christou A., 2018, return, making it possible to retrieve to the Earth some ma- MNRAS Letters 481, L49 Cellino A. et al., 2014, MNRAS 455, 2091 terials which might date back to the very early epochs of Cellino A., Belskaya I. N., Gil-Hutton R., 2015, Polarimetry planetary growth. of Stars and Planetary Systems (L. Kolokolova, J. Hought, A.-C. Levasseur Regourd, Eds.), 360 In most general terms, the number of existing families Cellino A. et al., 2014, MNRAS Letters, 439, L75 of Barbarians and the sizes of their parent bodies put criti- Cellino A., Belskaya I. N., Bendjoya P., di Martino M., Gil- cal constraints on the interpretation of the physical proper- Hutton R., Muinonen K., Tedesco E. F., 2006, Icarus, 180, ties and the possible origin of these objects (Devog`eleet al. 565 2018). In particular, the number of Barbarian families may Clemett S.J., Nakamura-Messenger K., McKay D. S., Sand- put some constraints on the initial inventory of these bodies, ford S. A., 2007, Lunar and Planet. Inst. Technical Report having been subject, since the end of the era of Late Heavy 38, 2091 Bombardment, to the general process of collisional evolu- Crovisier J. et al., 1997, Science 275, 1904 tion of the asteroid population, that can be investigated by Delb`oM., Walsh, K., Bolin, B., Avdellidou, C., Morbidelli, means of numerical models (Bottke et al. 2015). A., 2017, Science, 357, 1026 DeMeo F. E., Binzel R. P., Slivan S. M. and Bus S. J., There are many reasons to conclude that the study of 2009, Icarus, 202, 160 the Barbarian asteroids is an important topic in modern Devog`eleM., et al., 2018, Icarus, 304, 31 planetary science, and can shed some light on the processes Durda D. D. et al., 2007, Icarus, 186, 498 of early formation and subsequent evolution of planetesi- Farinella P. et al., 1998, Icarus, 132, 378 mals in our solar system. Further investigations are needed Giorgini P. et al., 1996, Bulletin of the American Astro- to solve some of the open questions concerning these inter- nomical Society, Vol 28, No. 3, p. 1158 esting objects. This is also another example of the impor- Gomes R. et al., 2005, Nature, 435, 466 tance of polarimetric studies of asteroids. This discipline has Hanner M. S. and Bradley J. P., 2004, Comets II, M. C. been somehow underexploited for a long time, until recent Festou, H. U. Keller, H. A. Weaver (eds.), University of years that have seen not only the discovery of Barbarians, Arizona Press, Tucson, p.555-564 but also the possibility to use polarimetric properties to in- Lisse C. M., et al., 2007, Icarus 187, 69 fer information about possible links between asteroids and Mainzer A. K. et al., NEOWISE Diameters and Albedos comets, as recently demonstrated in the case of polarimetric V1.0. EAR-A-COMPIL-5-NEOWISEDIAM-V1.0. NASA observations of (101955) Bennu (Cellino et al. 2018). Planetary Data System, 2016. Masiero J., Cellino A., 2009, Icarus, 199, 333 Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/stz451/5319141 by Washington University at St Louis user on 14 February 2019 Migliorini F., Zappal`aV., Vio R., Cellino A., 1995, Icarus, 118, 271 Milani A. et al., 2017, Icarus, 288, 240 Milani A. et al., 2014, Icarus, 239, 46 Morbidelli A. et al., 2005, Nature, 435, 462 Novakovi´cB., Cellino A., Kneˇzevi´cZ., 2011, Icarus, 216, 69 Rubincam D. P., 1995, J. Geophysical Res., 103, 1725 Schultze H., Kissel J., Jessberger E. K., 1997, Astronon. Soc. of the Pacific Conf. Series 122, 397 Spoto F., Milani A., Kneˇzevi´cZ., 2015 Icarus, 257, 275 Sunshine J. M., Connolly H. C., McCoy T. J., Bus S. J., La Croix L. M., 2008, Science, 320, 514 Tanga P., Cellino A., Michel P., Zappal`aV., Paolicchi P., Dell’Oro A., 1999 Icarus, 141, 65 Walsh K. J. et al., 2012, Meteoritics & Planet. Sci., 47, 1941 Walsh K. J. et al., 2011, Nature, 475, 206